assets/evals/fixtures/jtbd/gold.md
1 # Gold Mapping — JTBD switch interview (fixture)
2
3 Interview: co-owner of a coffee shop chain (3 locations, Krasnoyarsk), switching from manual tracking (Excel + WhatsApp) to the "Shift+" service (shift tracking and barista payroll calculation).
4 Alternatives considered: "Shift+" (chosen), "Time-Café" (finalist), "Chartline" (eliminated early — no payroll calculation).
5 Transcript: transcript.txt (93 lines, physical numbering, 1-based).
6
7 ---
8
9 ## Layer 1 — operational (facts + quote + line)
10
11 ### J1 — Core job (job statement)
12 Fact: "When I have several locations and shift/payroll tracking is done manually in Excel and WhatsApp, I want the timesheet and payroll calculation to be collected automatically factoring in rates and bonuses, so that I don't make mistakes in paying employees and don't have to be a schedule dispatcher."
13 Quote: "We wanted the shift timesheet to be collected automatically, for payroll to be calculated automatically factoring in the rate and holiday bonuses" (line 21)
14 Additional quote (part of the job statement about removing manual communication): "For payroll to be calculated without errors and for me not to be a schedule dispatcher in DMs" (line 67)
15
16 ### J2 — Trigger / struggling moment
17 Fact: a specific incident — underpaying barista Kate for two shifts due to a missed message in the WhatsApp chat; the respondent felt shame and immediately wrote to his partner that the system needed to change.
18 Quote: "There was a specific case — we underpaid one barista, Kate, for two shifts, because she logged them in WhatsApp and I just didn't scroll far enough in the chat and missed it" (line 9)
19 Additional quote (emotional reaction as part of the trigger): "I felt genuinely ashamed, honestly, because the person had worked, and I basically almost shortchanged her pay just because of the mess" (line 12)
20 Additional quote (immediate action after the trigger): "That's when I wrote to my partner that we needed to do something about this, we couldn't keep going like that" (line 13)
21
22 ### J3 — Job map (where it stalls)
23 Fact: the chain — learn about services (googling, asking in the owners' chat) → watch demos of several options → compare by criteria → choose between finalists → sign a contract → roll out location by location → deal with startup difficulties. Stalled at the "prepare/execute" stage: some baristas got confused with the new app in the first week, and the night shift was miscalculated once.
24 Quote (search and first demos): "We started googling, asked in a chat for coffee shop owners, they threw us like five services" (line 18)
25 Additional quote (rollout one location at a time): "About two weeks, I think, we onboarded one location at a time, so as not to break everything at once" (line 59)
26 Additional quote (friction at rollout start): "Two of the older baristas genuinely got confused the first week, called me with questions ten times a day" (line 61)
27 Additional quote (execution failure): "Also the night shift didn't calculate correctly right away at one location, we had to fix it manually once" (line 62)
28
29 ### J4 — Current solution and workarounds
30 Fact: before "Shift+" — three years of Excel plus WhatsApp shift lists; tried a Google Sheet with formulas (built by the partner himself), which "fell apart" when scaling to a third location; tried a shifts module in a checklist app, which baristas abandoned due to an overloaded interface.
31 Quote (baseline solution): "Well look, we ran everything in Excel for like three years plus shift lists in WhatsApp, baristas would send photos of the timesheet" (line 4)
32 Additional quote (Google Sheet abandoned): "We tried a Google Sheet with formulas for a while, my partner made it himself, but it fell apart on us when we added the third location, the formulas broke" (line 26)
33 Additional quote (checklist app abandoned, with reason): "Baristas complained that there were five screens just to check in for a shift, they eventually just stopped bothering, went back to WhatsApp anyway" (line 29)
34
35 ### J5 — Desired outcomes
36 Fact: payroll is calculated without errors; the respondent no longer performs the function of a manual schedule dispatcher; baristas see their schedule on their own phone; employees trust the calculations.
37 Quote: "We also wanted baristas to see their own schedule on their phone, instead of asking me in DMs for the tenth time" (line 22)
38 Additional quote (trust-in-calculations criterion): "I also wanted baristas to trust the calculations, so there wouldn't be a feeling like they were being cheated" (line 68)
39
40 ### J6 — Adjacent and emotional jobs
41 Fact: vacation planning got solved along the way (previously kept in a notebook); the respondent got emotional relief — stopped fearing he'd miscalculate someone's pay again, and didn't want to look like a dishonest employer in the baristas' eyes.
42 Quote (adjacent job — vacations): "Yeah, actually, planning vacations got simpler, before it used to be a separate nightmare in a notebook" (line 70)
43 Additional quote (emotional layer — relief and guilt): "And also, honestly, I started sleeping better, because I stopped being afraid that I'd messed up someone's pay again somewhere" (line 71)
44 Additional quote (emotional layer — reputation toward employees): "Yes, exactly, I didn't want to look, in the baristas' eyes, like an owner who shortchanges them out of carelessness" (line 73)
45
46 ---
47
48 ## Layer 2 — analytical (CAPITALIZED conclusion + anchor quote)
49
50 ### A1 — Forces of progress (unstable)
51 Conclusion: PUSH — recurring errors in manual tracking (underpaying Kate plus a similar case with another barista) and the respondent's lost time (3-4 hours a week on reconciliation). PULL — the vivid "Shift+" demo with automatic calculation and geolocation check-in, shown live. ANXIETY — fear that older employees wouldn't master the app, and fear of a repeat mistake during the transition itself. HABIT — WhatsApp is simple and requires no training, the partner suggested tolerating the old setup for another six months.
51 Anchor quote (push, final reassessment): "After that he agreed we couldn't keep dragging it out, the risk of errors was more costly than the habit" (line 55)
52 Additional anchor quote (habit): "My partner even said, maybe forget it, let's stick with the old system for another six months, why break everything now" (line 52)
53 Additional anchor quote (anxiety): "Of course there was, I was afraid the baristas would refuse to bother with a new app, we've got people of different ages, some of them aren't into technology at all" (line 48)
54 ⚑ Why unstable: the respondent himself gives an unverifiable quantitative estimate of the push/anxiety ratio ("forty percent doubt, sixty percent that we needed to act", line 75) and immediately qualifies it by noting that in hindsight there seems to have been less doubt (line 76) — the weight distribution between the forces requires the advice of independent runs (S3), it cannot be pinned down unambiguously.
55
56 ### A2 — Key unmet outcome
57 Conclusion: LACK OF EMPLOYEE TRUST IN PAYROLL ACCURACY — the gap is greatest between importance ("don't cheat employees") and actual satisfaction under the old setup (a real underpayment discovered after the fact).
58 Anchor quote: "I also wanted baristas to trust the calculations, so there wouldn't be a feeling like they were being cheated" (line 68)
59
60 ### A3 — "Hire/fire" criterion
61 Conclusion: DECIDED BY THE LIVE DEMO WITH AUTOMATIC CALCULATION AND GEOLOCATION CHECK-IN, SHOWN IMMEDIATELY AND VIVIDLY — not the stated price ("Time-Café" was 20% cheaper, but product access was only sent a week later and not "live," only on slides).
62 Anchor quote: "If I'm being completely honest, probably the live demo with the geolocation check-in and night shift calculation, that was the exact moment I understood" (line 87)
63
64 ### A4 — Segment by job
65 Conclusion: OWNER OF A SMALL MULTI-LOCATION FOOD SERVICE BUSINESS WHO MANUALLY PERFORMS THE JOB OF "SCHEDULE DISPATCHER AND PAYROLL CALCULATOR" AND WANTS TO AUTOMATE THIS ROLE AS THE NUMBER OF LOCATIONS GROWS — the job is defined not by demographics (coffee shop) but by the manual process scaling beyond a single location.
66 Anchor quote: "Yes, exactly, so I wouldn't be a schedule secretary, I don't have the time for that as it is" (line 24)
67
68 ### A5 — Adoption barrier
69 Conclusion: AGE GAP IN APP ADOPTION AMONG PART OF THE STAFF — confirmed in practice during the first week of rollout, even though at the selection stage it was only anxiety.
70 Anchor quote: "Two of the older baristas genuinely got confused the first week, called me with questions ten times a day" (line 61)
71
72 ### A6 — Key insight
73 Conclusion: THE CUSTOMER "HIRES" "SHIFT+" FOR [automatic and trustworthy shift and payroll tracking amid location-chain growth], BUT [habit toward WhatsApp and anxiety about older employees slowed the decision, until a specific underpayment incident made the risk of error more costly than the comfort of the old setup], SO [the product, for this segment, should emphasize demonstrating automatic calculation live and the simplicity of the first touch for older employees, rather than price].
74 Anchor quote: "But back then it seemed like it'd hold up fine as is, until it all blew up with Kate" (line 91)
75
76 ---
77
78 ## Intentionally ambiguous spots (for testing unstable cells)
79
80 1. Lines 74–76: the respondent gives a numeric estimate of the push/anxiety ratio ("forty percent doubt, sixty percent that we needed to act"), but immediately admits himself that memory distorts the picture in hindsight ("there wasn't that much doubt at all") — this touches A1, cannot be coded as a precise distribution of forces.
81 2. Lines 77–79: the respondent isn't sure whether the decision would have changed if "Time-Café" had sent access right away ("maybe, yeah, honestly I don't know, hard to say for sure, that's just guessing") — this touches A3, weakening the unambiguity of the conclusion about the decisive criterion, although the conclusion itself (the demo) is confirmed separately in line 87.
82 3. Trigger and signing dates: the respondent gets confused about dates twice — "last fall" is qualified as "maybe already November" (lines 5–7), and the contract signing — "I think it was in January, but I might be mixing it up, maybe it was already early February" (line 57). This is a normal feature of remembering a timeline, not a transcription error — record it as date uncertainty, don't guess the exact month.
83